46
7.4.
Paper VI
Rutting is one of the main distress modes in flexible pavements.
In thin pavement
structures, rutting is often associated with accumulation of permanent deformation in
unbound layers under repeated loads. Realistic prediction of surface rutting requires
models that can reliably capture the cumulative plastic
deformation of pavement
unbound layers under repetitive traffic loads. Paper VI presents an evaluation of several
models that incorporate the time-hardening concept for prediction of permanent
deformation of unbound materials using data from tests
conducted on two different
silty sand subgrade materials.
The time-hardening concept was adopted on the Tseng and Lytton (1989), Gidel et al.
(2001) and the Korkiala-Tanttu (2005) permanent deformation models so that the data
from the multistage RLT tests could be used for the evaluation. These permanent
deformation models combine the influence of the number of load applications with the
effect of material stress state. The details of these modified models are presented below.
The modified Tseng and Lytton (1989) model assumes a direct relationship between the
permanent strain and the resilient strain and is presented as follows:
)
(
0
1
ˆ
eq
i
i
i
N
N
N
r
i
p
e
N
[9]
where,
i
p
ˆ
is the accumulated permanent deformation and
i
r
is the resilient strain in
the
stress path
i
. The
0
,
and
are material parameters and
eq
i
N
is
defined as
follows:
1
1
ˆ
ln
o
r
p
eq
i
i
i
N
[10]
The modified Gidel et al. (2001) model additionally requires shear strength parameters
(obtained from static triaxial tests) and is rewritten as followed:
1
max,
max,
max,
max,
1
0
100
1
ˆ
i
i
i
u
a
i
B
eq
i
i
p
p
q
p
s
m
p
L
N
N
N
N
i
[11]
where
0
,
B
and
u
are material parameters.
max
p
and
max
q
are the maximum applied
hydrostatic stress and deviator stress, respectively, and
max
max
max,
q
p
L
i
.
The
parameters
m
and
s
are the slope and the intercept of the Mohr-Coulomb failure line
in the
q
p
space, respectively, obtained from static triaxial tests.
a
p
is
the reference
stress, here selected as100 kPa.
47
B
i
i
i
u
a
i
p
eq
i
p
q
p
s
m
p
L
N
i
1
1
max,
max,
max,
max,
0
1
ˆ
1
100
[12]
The modified Korkiala-Tanttu (2005) model for multistage test procedure is presented
as follows:
i
i
b
eq
i
i
p
R
A
R
N
N
N
C
N
i
)
(
ˆ
1
[13]
b
i
i
p
eq
i
CR
R
A
N
i
1
)
(
ˆ
1
[14]
where
C
is material parameter and
i
R
is the shear stress ratio of the deviator stress in
the stress path
i
to the deviator stress at failure.
A
is the maximum theoretical value
for the shear stress ratio and parameter
b
is recommended as follows (Korkiala-Tanttu,
2005):
d
cR
b
[15]
where
c
and
d
are material parameters.
In all of the equations above, the subscript
i
refers to the
th
i
stress path
.
The multistage
repeated load triaxial (RLT) tests were carried out on two silty sand subgrades at four
different moisture contents.
The test data were then used to optimize the material parameters
for each predictive
model and moisture content. The calibrated model curves and the measured data
together with the shakedown ranges are presented in Figures 31 and 32.